New powerful drugs for treatment and prevention of osteoporosis are already available or are currently under development. The primary endpoint for these studies continues to be fracture incidence, the secondary endpoint typically is bone mineral density (BMD), both of which are fraught with problems. The former requires a very large number of study subjects as fractures are relatively rare events and require long observation periods, therefore resulting in excessive costs and long development cycles. BMD has been an unreliable indicator of treatment efficacy showing often disproportionately small increases relative to the extent of fracture reduction caused by the drug, in contrast to the much larger architectural changes that occur in the trabecular bone (TB) network. Progress in 3D high-resolution MRI ([unreadable]-MRI) now allows acquisition of images from which the topology of the TB network can be established. Nevertheless, in spite of these advances, structure plays only a surrogate role and it is not known which parameters and combinations thereof are optimal in terms of responding to treatment, and which are most representative of strength. Advances in micromechanical modeling now permit micro finite-element ([unreadable]-FE) computations of TB mechanical competence, potentially providing insight into the mechanical implications of disease progression and regression. We have, in preliminary work, examined the feasibility of quantifying the effect of antiresorptive treatment in a small cohort of patients and demonstrated significant improvement in the elastic moduli after computing the full stiffness matrix on the basis of MR images acquired in the distal tibia. While encouraging, the feasibility of deriving mechanical parameters from in vivo images as possible end points in clinical trials demands further scrutiny. In this project we advance the hypothesis that the treatment-induced changes in the bone's mechanical parameters, estimated from [unreadable]-FE calculations on the basis of in vivo [unreadable]-MRI, represent a quantitative measure of treatment response. The proposed research, involving four specific aims, seeks to (1) further develop algorithms for processing in vivo [unreadable]-MRI data with improved motion correction and serial registration capabilities; (2) evaluate the effect of resolution and noise on the derived mechanical indices in images of intact specimen under conditions of in vivo [unreadable]-MRI as well as by simulation previously or currently in progress to determine the effect of treatment and comparing the mechanical with of specimen [unreadable]-CT images; (4) apply [unreadable]-FE analysis to three longitudinal [unreadable]-MRI studies performed structural indices. [unreadable] [unreadable] [unreadable]